luck factor, high hashrate in hive OS will not guarantee getting many shares
Thanks for this update buddy!
Mine is stable at @1090 but it’s very inefficient compared to Samsung:
7W more is unacceptable being the same chipset and I also had to undervolt Micron memory voltage as you can see, Samsung is at default @1350.
Hynix and Micron = crap
1 Like
this is shown in hiveos …you only can compare real stats …do you have yours?
doesnt seems bad I have a powercolor hellhound that cant go better then 31,6MH/s and 55W (Samsung chip)
it stucks on 1105 on memory and thats a huge issue on the samsung chip.
Of course I tested every single card at the wall, real hashrate on the pool and the share per minute (24h). I have 12 MSI cards with Samsung memory, all from 47w to 49w and 32.5 to 34 Mh/s (it depends on the miner), at 48w they produce a little better than Sapphire at 56w (at @1150 frequency Samsung memory is still stable).
How did you guys test the power consumption from wall? I want to do that with my cards.
buy a power tester? 
I’m using two ASUS RX 6600XT Dual OC on Kernel 5.10.0-hiveos #72 and OS 0.6-212@211211 (Latest)
Both are detected but only GPU1 will hash at 28MH @ 52W while GPU0 will not hash even though all temps and fan speeds are recorded for it. If i only plug in GPU0, it will hash at 28MH but at 110W (confirmed by power meter).
I have no clue, but both were out of box brand new and never OC-ed. I don’t see many people using ASUS 6600XTs. I dont have a third one yet to test if using more than one ASUS GPUs causes issues. I ran GPU0 on FurMark stress test on another computer for 15 mins at 100% and it ran flawless.
I’m so lost… Any recommendations? (I’m using GPURisers and cables, all working on TZ590-BTC DUO BIOSTAR Group)
oh jear, greate idea… are there some specific ones for GPUs? Or just measure up without GPUs and then with the GPUs?
I have only one on the wall for the whole rig. In germany around 10 Euro for one thing.
So you can Test your power without cards and then add one card after the other and look for change. every card has more power for itself, the riser etc …
what does it help you you know your card uses 50 watts, butt you have more power for the riser and for example psu efficience level etc… look at live change on adding every card
1 Like
nah, @57W? i don’t think so…
1 Like
okay, also stumpfes messen des servers mit und ohne gpu. dachte vielleicht gibt ein “spezielles” messgerät.
Nicht das ich wüsste - du müsstest jede Karte einzeln messen.
Mach aber, wie gesagt, auch mehr Sinn so … weil eben die Riser, die PSU Effizienz usw. alles viel mehr Strom fressen, als man an sich meint Wenn du dann ne 60Watt Karte anschließt und siehst 100Watt flöten gehen, staunt man schonmal nicht schlecht
Aktuell hab ich am Gerät 2 Rigs mit 630 und 646 Watt Verbrauch lt. Hiveos.
Reell gehen da etwa 1555 Watt durch. Und ich nutze EVGA 1000GQ Netzteile mit 94% Effizienz.
Ich nutze stinknormale Wattmesser von Pearl, gibts zeitweise für 7-8EUR. Im Endeffekt interessiert es mich am meisten, was das gesamte Rig verbraucht.
Mir ist jedoch aufgefallen, dass die Verbräuche nicht skalierbar bzw. linear sind - mit jeder neuen Grafikkarte verbraucht auch dein Mainboard anders.
Ich wollte da auch keine Wissenschaftsarbeit draus machen wegen ein paar Watts hin oder her.
Wer hier jede GPU anders takten will um das letzte raus zu holen, wir mit der Angabe “Watt an der Steckdose” nicht so ganz zufrieden sein.
@Lalading @Kujko
Ihr habt schon recht, man sollte da keine allzugroße Wissenschaft drauß machen.
Meine Werte sahen folgendermaßen aus:
Server ohne GPUs (ohne HiveOS VM): 40W
Server +2xGPUs +HiveOS VM: 160W
Server +3xGPUs +HiveOS VM: 225W
(alle GPUs sind 6600xt)
Denke das hier auch ca 10W die Virtualisierung der VM ausmacht. Finde die Werte aber letztendlich sehr gut für ca 95MH/s.
Is this card not supported by Hiveos? I’ve had 8-10 running previously, but not for very long and it seems to freeze hive and less cards work each time. I’m down to 4 cards on a rig that was previously running 10 with 2 nvidia cards as well. It looks like riser issues, but I’ve never had this many issues with risers and I completely swapped out the entired set of risers as well. Also tried on the motherboard that skips the pcie slot and goes straight to the usb to remove those as possible causes of errors. Booted 10 cards the first 2 times no problem, then on 3rd boot it stopped posting. Removed all cards and seems like mb is dead now. Do we need to use windows to mine with these cards? It’s been nothing but a headache and I have had many of these cards sitting around because I haven’t heard a solid answer from anyone yet regarding these cards.
HiveOS Work defenitly with 6600xt. I run three of them in a VM without problems since they released.
1 Like
Seems very problematic tho. I can run a handful with no problems, but normally I run 8+ cards per rig. With 8+ cards, it seems that my hardware stops working. Not sure why. I swapped out all risers/psu/ram/mobo and now it will only hash with 3 cards, the rest don’t even show up. Never had so much trouble mining on any other cards. It seems like it’s killing my hardware. I see risers getting a lot of blame, but I’ve never had this many bad risers ever. In a 14 card rig, the same risers have run for months now with no issues. Same with two other 8 card rigs. It’s very unlikely that the hardware is the issue.
Would be great if there was some kind of update from hive ppl regarding this. It shouldn’t kill my hardware like this.